diff --git a/examples/andriod_assistant/actions/screenshot_parse.py b/examples/andriod_assistant/actions/screenshot_parse.py
index 1dbd45b5e..ef9454979 100644
--- a/examples/andriod_assistant/actions/screenshot_parse.py
+++ b/examples/andriod_assistant/actions/screenshot_parse.py
@@ -104,7 +104,7 @@ class ScreenshotParse(Action):
                 return
         elif isinstance(op_param, SwipeOp):
             x, y = elem_bbox_to_xy(elem_list[op_param.area - 1].bbox)
-            res = env.step(EnvAPIAbstract("user_swipe", kwargs={"x": x, "y": y, "dir": op_param.swipe_orient, "dist": op_param.dist}))
+            res = env.step(EnvAPIAbstract("user_swipe", kwargs={"x": x, "y": y, "orient": op_param.swipe_orient, "dist": op_param.dist}))
             if res == ADB_EXEC_FAIL:
                 # TODO
                 return
diff --git a/metagpt/environment/android_env/android_env.py b/metagpt/environment/android_env/android_env.py
index 87b49750d..6c83be530 100644
--- a/metagpt/environment/android_env/android_env.py
+++ b/metagpt/environment/android_env/android_env.py
@@ -2,8 +2,11 @@
 # -*- coding: utf-8 -*-
 # @Desc   : MG Android Env
 
+from pydantic import Field
+
 from metagpt.environment.android_env.android_ext_env import AndroidExtEnv
 
 
 class AndroidEnv(AndroidExtEnv):
-    pass
+    rows: int = Field(default=0, description="rows of a grid on the screenshot")
+    cols: int = Field(default=0, description="cols of a grid on the screenshot")
diff --git a/metagpt/environment/android_env/android_ext_env.py b/metagpt/environment/android_env/android_ext_env.py
index f5e62eb5e..7467d394c 100644
--- a/metagpt/environment/android_env/android_ext_env.py
+++ b/metagpt/environment/android_env/android_ext_env.py
@@ -67,39 +67,39 @@ class AndroidExtEnv(ExtEnv):
         return devices
 
     @mark_as_readable
-    def get_screenshot(self, ss_name: str, local_save_dir: Path) -> str:
+    def get_screenshot(self, ss_name: str, local_save_dir: Path) -> Path:
         """
         ss_name: screenshot file name
         local_save_dir: local dir to store image from virtual machine
         """
         assert self.screenshot_dir
-        ss_remote_path = str(Path(self.screenshot_dir).joinpath(f"{ss_name}.png"))
+        ss_remote_path = Path(self.screenshot_dir).joinpath(f"{ss_name}.png")
         ss_cmd = f"{self.adb_prefix_shell} screencap -p {ss_remote_path}"
         ss_res = self.execute_adb_with_cmd(ss_cmd)
 
         res = ADB_EXEC_FAIL
         if ss_res != ADB_EXEC_FAIL:
-            ss_local_path = str(Path(local_save_dir).joinpath(f"{ss_name}.png"))
+            ss_local_path = Path(local_save_dir).joinpath(f"{ss_name}.png")
             pull_cmd = f"{self.adb_prefix} pull {ss_remote_path} {ss_local_path}"
             pull_res = self.execute_adb_with_cmd(pull_cmd)
             if pull_res != ADB_EXEC_FAIL:
                 res = ss_local_path
-        return res
+        return Path(res)
 
     @mark_as_readable
-    def get_xml(self, xml_name: str, local_save_dir: Path) -> str:
-        xml_remote_path = str(Path(self.xml_dir).joinpath(f"{xml_name}.xml"))
+    def get_xml(self, xml_name: str, local_save_dir: Path) -> Path:
+        xml_remote_path = Path(self.xml_dir).joinpath(f"{xml_name}.xml")
         dump_cmd = f"{self.adb_prefix_shell} uiautomator dump {xml_remote_path}"
         xml_res = self.execute_adb_with_cmd(dump_cmd)
 
         res = ADB_EXEC_FAIL
         if xml_res != ADB_EXEC_FAIL:
-            xml_local_path = str(Path(local_save_dir).joinpath(f"{xml_name}.xml"))
+            xml_local_path = Path(local_save_dir).joinpath(f"{xml_name}.xml")
             pull_cmd = f"{self.adb_prefix} pull {xml_remote_path} {xml_local_path}"
             pull_res = self.execute_adb_with_cmd(pull_cmd)
             if pull_res != ADB_EXEC_FAIL:
                 res = xml_local_path
-        return res
+        return Path(res)
 
     @mark_as_writeable
     def system_back(self) -> str:
@@ -127,20 +127,20 @@ class AndroidExtEnv(ExtEnv):
         return press_res
 
     @mark_as_writeable
-    def user_swipe(self, x: int, y: int, dir: str = "up", dist: str = "medium", if_quick: bool = False) -> str:
+    def user_swipe(self, x: int, y: int, orient: str = "up", dist: str = "medium", if_quick: bool = False) -> str:
         dist_unit = int(self.width / 10)
         if dist == "long":
             dist_unit *= 3
         elif dist == "medium":
             dist_unit *= 2
 
-        if dir == "up":
+        if orient == "up":
             offset = 0, -2 * dist_unit
-        elif dir == "down":
+        elif orient == "down":
             offset = 0, 2 * dist_unit
-        elif dir == "left":
+        elif orient == "left":
             offset = -1 * dist_unit, 0
-        elif dir == "right":
+        elif orient == "right":
             offset = dist_unit, 0
         else:
             return ADB_EXEC_FAIL